Biosynthesis of the Alkaloids of Chelidonium

administered to Chelidonium majus plants the benzo [c] phenanthridine alkaloids, chelidonine and sanguinarine, became labeled. By systematic degradati...
1 downloads 0 Views 439KB Size
INCORPORATION OF TYROSINE INTO CHELIDONINB

Feb. 20, 1963 [CONTRIBUTION FROM

THE SCHOOL OF

473

CHEMISTRY, UNIVERSITY OF MINNESOTA, MINNEAPOLIS, MI".]

Biosynthesis of the Alkaloids of Chelidonium majus. I. The Incorporation of Tyrosine into Chelidoninel BY EDWARD LEETE~ RECEIVED AUGUST20, 1962 When ~ ~ - t y r o s i n e - 2 - C was 1 ~ administered t o Chelidonium majus plants the benzo[c]phenanthridine alkaloids, chelidonine and sanguinarine, became labeled. By systematic degradation of the radioactive chelidonine it was established that the label was located at specific positions, supporting the hypothesis that the biosynthesis of this class of alkaloid involves two molecules of tyrosine or closely related metabolites.

I t has been suggested3m4that the benzo[c]phenanthridine alkaloids, of which chelidonine (111)and sanguinarine (VI) are members, are formed by a modification of a berberine skeleton (11) indicated schematically in Fig. 1. It is generally accepted that the berberine alkaloids are formed from two molecules of tyrosine or its hydroxylated derivative, 3,4-dihydroxyphenylalanine, and a one carbon f r a g ~ n e n t . ~Quite recently this hypothesis has been strongly supported by tracer experiments. Spenser and Gear6 fed tyrosine-2-C1* (I) to Hydrastis canadensis plants and obtained radioactive berberine which was labeled solely a t the expected positions, C1 and Cz. If chelidonine is indeed formed by a rearrangement of a berberine skeleton we would therefore expect chelidonine derived from tyro~ i n e - 2 - C 'to ~ be labeled a t positions C4b and C,, which

posed that two molecules of 3,4-dihydroxyphenylacetaldehyde (IV), which is considered to arise by the oxidative decarboxylation of 3,4-dihydroxyphenylalanine, condense together to give the aldol (V), from which the nucleus of chelidonine and related alkaloids can be formed by two Mannich reactions involving ammonia and formaldehyde. Wenkert8 has adumbrated a similar scheme involving two molecules of prephenic acid instead of the hydroxylated phenylacetaldehyde. m'e have tested these hypotheses by feeding DLtyrosine-2-C14 to five-month old Chelidonzum majus plants growing in hydroponics. The tracer was absorbed rapidly by the roots and after eight days the plants were harvested yielding radioactive chelidonine and sanguinarine, which were purified by chromatography on alumina and Florid. The radioactive chelidonine, which had a constant specific activity on repeated crystallization, was degraded according to the scheme illustrated in Fig. 2. Heating with hy-

J& Hoocw) COOH

"

r foH

111-

I OH

P O

I?'

I COOH ('OOH \ 11

+

4 Ho@

0

Yiil

J. Y O

EHO

t

XI Fig. 2.-Degradative

Fig. 1.-Biogenetic

schemes for the benzo[c]phenanthridine alkaloids.

are indicated by heavy dots in formula 111. I t should be mentioned that the alternate scheme of Manske' would lead to the same pattern of labeling. He pro(1) An account of this work was presented a t the Second International Symposium on t h e Chemistry of Natural Products, Prague, August 27 t o September 2, 1962. This investigation was supported by a research grant, CY-5336, from the National Institutes of Heaith, U. S. Public Health Service. (2) Alfred P. Sloan Research Fellow, 1962-1964. (3) R . Robinson, "The Structural Relations of S a t u r a l Products," Clarendon Press, Oxford, 1955, p. 89. (4) R. E.Turner and R. B. Woodward i n "The Alkaloids," Vol. 111, Ed., R. H . F. Manske and H. L. Holmes, Academic Press, New York, N. Y., 1953, p. 57. (5) Cf.ref. 3, p. 86. (6) I. D. Spenser and J. R. Gear, Proc. Chem. Soc., 228 (1962). (7) R. H. F. Manske in "The Alkaloids," Vol. IV, Ed. R. H. F. Manske and H . L. Holmes, Academic Press, New York, N.Y., 1954, p. 5.

XI1 scheme for the chelidonine-C14.

driodic acid yielded methyl iodide which was absorbed in triethylamine to give triethylmethylammonium iodide which had negligible activity, implying that there was no activity on the N-methyl group. Oxidation with potassium permanganate according to the procedure of Spath and Kuffnerg yielded a mixture of hydrastic (VII) and 3,4-methylenedioxyphthalic acid (VIII) which were converted to their N-ethylimides and separated by chromatography. Between them, these two compounds contain all the carbons of chelidonine except the N-methyl group and CI1. Since the Y-methyl group was inactive, activity a t CI1 could be determined by difference and i t was found that this position contained 61% of the total activity of the chelidonine. This method of determining the activity a t CI1 is dependent on reliable values for the specific activities of chelidonine and the N-ethylphthalimides. Subsequent degradation confirmed the activity of the latter compounds, and several derivatives of chelido(8) E. Wenkert, Experientio, 16, 165 (1959). (9) E . Spath and F. Kuffner, Ber., 64, 370 (1931).

EDWARD LEETE

474

nine were prepared, all of which had the same specific activity. Attempts were made to determine activity at CH directly by oxidation to a Cll ketone which could then be phenylated and subsequently oxidized to benzoic acid. However several different methods of oxidation failed to yield the desired ketone. The activity of the 3,4-methylenedioxy-N-ethylphthalimide (X) was negligible, a result which is con sistent with the hypotheses illustrated in Fig. 1. The N-ethylhydrastimide (IX) contained 39% of the activity of the chelidonine. It was hydrolyzed to hydrastic acid, dehydrated to the anhydride, and then converted to hydrastimide (XI) by heating with urea. Treatment of this imide with sodium hypochlorite yielded 4,5methylenedioxyanthranilic acid (XII) which had half the specific activity of the hydrastimide. This result indicates that all the activity was located on the carbonyl groups of the hydrastimide. Because of the symmetry of this compound the activity could be located on one or both of the carbonyl groups. We can thus state that the chelidonine was labeled a t C4b or Clz or both. However, it is impossible to conceive of a biogenetic scheme by which tyrosine-2-C14 could yield chelidonine labeled a t Clz and we feel confident that the alkaloid was labeled only a t C4b and Cll. The activity was not equally divided between these two positions, and this result is another demonstration of the differential utilization of a single precursor in the biosynthesis of two segments of a "dimeric" alkaloid. Other examples of this phenomenon have been discovered by Gear and SpenserlO in their study of the biosynthesis of hydrastine from tyrosine-2-C14, and by Rapoport and co-workers'l studying the incorporation of carbon d i 0 ~ i d e - C into ~ ~morphine. Our results are consistent with the formation of chelidonine by the rearrangement of a berberine skeleton, but do not support Manske's hypothesis' which requires equal labeling at C t b and C1l. The coexistence of the berberine and benzo [clphenanthridine alkaloids in the Papaveraceae (Chelidonium ma&, Glaucium corniculatum,lZand several other species) and Rutuceae (Toddalia a c u l e a t ~ ~is~strong ) circumstantial evidence in favor of their biosynthesis from common precursors.

Experimental14 Administration of the ~ ~ - T y r o s i n e - Z - Cto~ ~ Chelidonium majus and Isolation of Chelidonine and Sanguniarine .-The Chelidonium majus plants were grown from seed in soil until they were about five months old. Four plants were then transferred t o a hydroponics setup in which the roots were placed in aerated tap water. After several days new roots mere produced and ~ ~ - t y r o s i n e S - Cl 6l *(78.2 mg. 2.28 X lo8 d.p.m.) was added to the aqueous solution in which the roots were growing. Uptake of the tracer was rapid and after three days less than 1% remained in the hydroponic solution. Eight days after administration of the tracer the plants were harvested (wet wt. 866 9.) and mascerated in a Waring blender with chloroform (3 1.) and 15 N ammonia (100 ml.). After standing for two days the mixture was filtered through cloth yielding an aqueous phase (530 ml., 2.0 X IO7 d.p.m.l8) and a chloroform layer which was taken t o dryness in a rotary evaporator. The residue (6.5.g.) was extracted five times with 100 ml. portions of hot 10% acetic acid. The deep orange solution was made basic with concentrated ammonia (120 ml.) and extracted with chloroform. (10) J. R. Gear and I. D. Spenser, Nature, 191, 1393 (1961). (11) H.Rapoport, N. Levy and F. R. Stermitr, J . A m . Chcm. Sac., 83, 4298 (1961). (12) J. Slavik and L. Slavikova, Collection Czech. Chcm. Commun; 22, 279 (1957). (13) T. R. Govindachari and B. S. Thyagarajan, J. Ckcm. Sac., 769 (1956). (14) Melting points are corrected, and analyses were carried out by Mrs. Olga Hamerston and her assistants at the University of Minnesota. (15) Purchased from Volk Radiochemical Co., Skokie, Illinois. (16) Activities were determined in a Nuclear Chicago Model C-115 low background Q gas flow counter. Determinations were carried out on samples of finite thickness making corrections for efficiency and self absorption.

Vol. 85

Evaporation of the dried chlorofonn extract yielded a mixture of crude alkaloids (2.2 g., 7.0 X IO6 d.p.m.). These alkaloids were dissolved in benzene and chromatographed on a column of Woelm alumina (600 g.) (Activity 11). Elution was carried out successively with benzene, methylene chloride, and finally with methylene chloride containing increasing amounts of ethanol (l-lO%). The composition of the eluents was determined by thin layer chromatography, using alumina on glass plates, with methylene chloride as the developing solvent. Crude chelidonine (741 mg.) was eluted with 1% ethanol in methylene chloride, and was rechromatographed on a column of Florisil eluting with methylene chloride containing increasing amounts of ethanol (0-500/0). Chelidonine was present in the fractions obtained by elution with 10% ethanol in methylene chloride, and was crystallized fiom aqueous methanol yielding colorless prisms of the monohydrate (297 mg., 1.13 X lo6 d.p.m./mM) m.p., 135-136', not depressed on admixture with an authentic specimen. Sanguinarine (95 mg., 1.1 X lo5 d.p.m./mM) was obtained from the more poJar fractions of the alumina and Florisi1 columns, and was crystallized from a mixture of ether and methanol. Work is proceeding on the isolation and characterisation of the minor alkaloids which have been obtained from the chromatographic columns. Degradation of the Chelid~nine-C*~.(a) Demethylation.Chelidonine (56 mg.) was mixed with ammonium iodide (50 mg.), gold chloride (1 mg.) and freshly distilled hydriodic acid (2 ml., d. 1.7) and heated to 360' in a current of nitrogen. The hydriodic acid which distilled was condensed in a water bath maintained a t 70-80'. The nitrogen stream then was passed through a solution containing 2.5% cadmium sulfate and 2.5% sodium thiosulfate, and then into a 5% solution of triethylamine in ethanol (5 ml.) cooled to -80'. After standing overnight the triethylamine solution was evaporated yielding triethylmethylammonium iodide (32 mg.) which, after crystallization from a mixture of ethanol and ether, was identical (mixed m.p., infrared spectrum) with an authentic specimen. ( b ) Oxidati~n.~-Chelidonine (250 mg.) was dissolved in 2 N sulfuric acid ( 5 1111.) and then diluted t o 150 ml. with water. Sodium carbonate solution was added until the mixture was turbid and then potassium permanganate (1.5 g.) in water (150 ml.) was added with stirring during 3 hr. The mixture was warmed on a steam bath for 30 min., then cooled, decolorized with sulfur dioxide, and finally evaporated t o small bulk (30 ml.). This solution was made strongly acidic with hydrochloric acid and extracted continuously for 2 days with ether. The residue obtained on evaporation of the ether extract was dissolved in dilute ammonia, and calcium chloride solution added when calcium oxalate precipitated. The filtered solution was acidified with hydrochloric acid and extracted again with ether. The residue from this second extraction was dissolved in 30y0 ethyl; amine solution ( 5 ml.), evaporated t o dryness, heated a t 180 and then sublimed (150", 0.001 mm.). The sublimate (42.5 mg.) was dissolved in a 1: 1 mixture of benzene and petroleum ether (b.p. 60-70') and chromatographed on Woelm alumina (Activity 111),eluting first with 1:1 benzenepet. ether, and then with pure benzene. The N-ethylphthalimides readily were detected on the column by their strong fluorescence in ultraviolet light. N-Ethylhydrastic acid (green fluorescence) (16.9 mg.) m.p. 166-167', was eluted first followed by 3,4methylenedioxyN-ethylphthalimide (blue fluorescence) (22.3 mg.) m.p. 124125.'. The identity of these imides was established by comparison with authentic specimens.17J8 (c) Hydrastimide.-The active N-ethylhydrasthide which had been diluted with inactive material to give a total wt. of 90 mg. was refluxed with 30% potassium hydroxide ( 5 ml.) for 30

TABLE I ACTIVITYOF CHELIDONINE AND ITS DEGRADATION PRODUCTS Activity (10-1 X d p m

/

mM )

Chelidonine (111) Chelidonine hydrochloride O-Acetyl~helidonine~~ Triethylmethylammonium iodide N-Ethylhydrastimide (IX) 3.4-Methylenediosy-N-ethylphthalimide (X) Hydrastimide (XI) 4,5-Methylenedioxynnthranilic acid (XII)

1.13 1.17 1.12